On 30/04/2019 16:06, Lorenzo Pieralisi wrote: > On Tue, Mar 12, 2019 at 06:34:55PM +0100, Marc Gonzalez wrote: > >> On 12/03/2019 18:18, Marc Gonzalez wrote: >> >>> On 12/03/2019 13:42, Stanimir Varbanov wrote: >>> >>>> I wonder, in case that dw_pcie_setup_rc() already has a write to >>>> PCI_CLASS_DEVICE configuration register to set it as a bridge do we >>>> still need to do the above fixup? >>> >>> I don't know, I don't have an affected device. Unless the msm8998 /is/ affected, >>> and dw_pcie_setup_rc() actually fixes it? >> >> I think you hit the nail on the head... >> >> If I comment out >> //dw_pcie_wr_own_conf(pp, PCI_CLASS_DEVICE, 2, PCI_CLASS_BRIDGE_PCI); >> from dw_pcie_setup_rc() >> then pci_class() returns 0xff000000 instead of 0x6040000 >> >> So perhaps you're right: the quirk can be omitted altogether. >> Unless it is not possible to program the device class on older chips? > > Marc, > > I would drop this patch from the PCI queue since in a different > form it was already merged, please let me know if I am wrong. I'm confused because you speak of *dropping* this patch (v1) -- but v1 was never picked up AFAIK. You picked up v5 on March 29: https://patchwork.kernel.org/patch/10869519/ I see it in linux-next as 915347f67d41857a514bed77053b212f3696e8a3 Will Bjorn send it to LT during the merge window for 5.2? Regards.
